Volume 22, Issue 8, Pages 1087-1097 (August 2015) The Prototypic Cyclotide Kalata B1 Has a Unique Mechanism of Entering Cells  Sónia Troeira Henriques, Yen-Hua Huang, Stephanie Chaousis, Marc-Antoine Sani, Aaron G. Poth, Frances Separovic, David J. Craik  Chemistry & Biology  Volume 22, Issue 8, Pages 1087-1097 (August 2015) DOI: 10.1016/j.chembiol.2015.07.012 Copyright © 2015 Elsevier Ltd Terms and Conditions

Chemistry & Biology 2015 22, 1087-1097DOI: (10. 1016/j. chembiol. 2015 Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 1 Structural Properties of Cyclotides (A) Three-dimensional structure (PDB: 1NB1, left) and sequence (right) of the prototypic cyclotide, kB1. The knotted arrangement of the disulfide bonds is shown in light gray and Cys residues are labeled I to VI. The segments between Cys are named loops 1–6. (B) Surface structure representation of kB1 in two views. Left: The hydrophobic residues (P3, V4, V10, W23, P24, and V25) of kB1 are exposed and form a patch at the surface of the molecule. Residues that are important for membrane binding localize to one side of the molecule, and include the residues in the hydrophobic patch and also those in the so-called bioactive face (G6, E7, T8, N15, T16, and R28). Right: Residues located in the “amendable” face (G1, G18, T20, S22, T27, and N29) are delineated with a line. They are named amendable because when replaced with a Lys residue they improve the membrane-binding properties of the peptide (Henriques et al., 2011; Huang et al., 2010). See also Table S1. Chemistry & Biology 2015 22, 1087-1097DOI: (10.1016/j.chembiol.2015.07.012) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 2 Internalization of kB1 (A–D) Internalization of labeled peptides into HeLa cells was followed by mean fluorescence emission intensity of Alexa Fluor 488 using flow cytometry (excitation at 488 nm and emission at 530/30 nm). Data points represent mean ± SD of three or more experiments; 10,000 cells were screened per sample. (A) Percentage of fluorescent cells obtained upon treatment with 4 μM TAT∗, TAT-G∗, T20K∗, or T16K∗ at 37°C at different incubation times and after addition of the non-permeable fluorescence quencher trypan blue (160 μg/ml, n = 3). (B) Mean cell fluorescence signal upon treatment with 4 μM TAT∗, TAT-G∗, T20K∗, or T16K∗ for 1 hr at 37°C before and after addition of trypan blue (TB). Data were normalized to the signal obtained with T20K∗ after trypan blue (n = 6). (C) Fluorescence histograms of HeLa cells obtained after incubation with 4 μM TAT∗, TAT-G∗, or T20K∗ for 1 hr at 37°C and after addition of trypan blue. (D) Mean fluorescence signal of HeLa cells obtained upon treatment with increasing concentrations of T20K∗, V25K∗, E7K∗, or T16K∗ for 1 hr at 37°C and after addition of trypan blue (n = 3). (E) Confocal live cell imaging micrographs obtained with HeLa cells after incubation with 5 μM T20K∗ or T16K∗ for 1 hr at 37°C. 1024 × 1024 pixels; scale bar, 10 μm. The bar graph shows the average cell fluorescence ± SD (n = 8 cells) in each micrograph. Values are normalized to the average of T20K∗. See also Figures S1 and S2. Chemistry & Biology 2015 22, 1087-1097DOI: (10.1016/j.chembiol.2015.07.012) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 3 The Involvement of Endocytosis in the Internalization of kB1 Mutants (A) Internalization of 4 μM T20K∗ into HeLa cells after 1 hr incubation at 4°C or at 37°C in the absence/presence of the endocytic inhibitors: Dynasore (80 μM), chlorpromazine (CPZ; 10 μg/ml, 28 μM), 5-(N-ethyl-N-isopropyl)-amiloride (EIPA; 10 μM), methyl-β-cyclodextrin (MβCD; 4 mM), and cytochalasin D (CytD; 20 μM), or the combination of Dynasore with EIPA, MβCD, or CytD. Internalization was followed by fluorescence emission of Alexa Fluor 488 using flow cytometry (excitation at 488 nm and emission at 530/30 nm). 10,000 cells were analyzed per sample and data points represent mean ± SD. Data were normalized to the mean fluorescence emission signal obtained with 4 μM T20K∗ after treatment with trypan blue (n = 6). Statistical analysis was performed by one-way ANOVA, with Bonferroni test (∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). (B and C) Mean fluorescence emission (excitation 488 nm, emission 530/30 nm) of HeLa cells incubated at 37°C at various incubation times with (B) 4 μM T20K∗ without/with 30 min pre-treatment with 50 μM Myr-PI, and (C) with 4 μM T20K-pHrodo without/with 30 min pre-treatment with 80 μM Dynasore or with 50 μM Myr-PI. Data points represent mean ± SD of three experiments, and were normalized to the average at 60 min without inhibitor; 10,000 cells were analyzed per sample. (D) Confocal live cell imaging micrographs obtained with HeLa cells upon incubation with 10 μM T20K-pHrodo at 37°C. 2048 × 2048 pixels; scale bar, 10 μm. The plot shows the average cell fluorescence integrated density (n = 9) ± SEM within the image field over time. Fluorescence integrated density was calculated using ImageJ. (E) SPR sensorgrams obtained upon injection of 64 μM kB1 over POPC/POPE (4:1) bilayers either in buffer at pH 7.4 (10 mM HEPES buffer containing 150 mM NaCl) to mimic membrane-binding properties in the extracellular and cytoplasmic pH, or in buffer at pH 5.5 (20 mM sodium acetate buffer containing 150 mM NaCl) to examine membrane-binding affinity in the endosome compartment environment. See also Figures S3 and S4. Chemistry & Biology 2015 22, 1087-1097DOI: (10.1016/j.chembiol.2015.07.012) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 4 Effect of kB1 on Lipid Membranes and the Role of PE Phospholipids (A) GVs labeled with 1% NBD-PE visualized by confocal microscopy after incubation with peptide (15 min): 1, POPC with 10 μM kB1 (512 × 512 pixels); 2, 3, POPC/POPE (4:1) with 5 μM kB1; 4, POPC/POPE (4:1) with 10 μM T16K (scale bar 10 μm; micrograph size is 512 × 512 pixels for 1, 2 and 4; 241 × 241 pixels for 3). (B) SPR sensorgram obtained upon injection of 40 μM kB1 over POPC or DPPC bilayers containing 20% of DPPE, DMPE, POPE, or DOPE deposited over an L1 chip surface. kB1 was injected for 180 s and dissociation followed for 600 s. Response units were converted into peptide-to-lipid ratio (P/L; mol/mol) to normalize the response to the amount of peptide as before (Henriques et al., 2011). (C) P/L at the end of peptide injection plotted versus peptide concentration. (D) Static (upper) and MAS (lower) 31P NMR spectra of POPC/DOPE (1:1) MLV without (black) and with (gray) kB1 at P/L 1:10. Insets show deconvolution of the MAS spectra using Lorentzian functions and their sum (dashed lines). Experiments were performed at 30°C. (E) Pake (upper) and dePaked (lower) 2H NMR spectra of POPC/d31-POPC/DOPE (1:1:2) without (black) and with (gray) kB1 at P/L 1:10. See also Table S2. Chemistry & Biology 2015 22, 1087-1097DOI: (10.1016/j.chembiol.2015.07.012) Copyright © 2015 Elsevier Ltd Terms and Conditions

Figure 5 Model of the Mechanism of Internalization of kB1 (1) Targeting of the cell membrane through PE phospholipids increases the concentration of kB1 in the membrane vicinity; (2) insertion into the outer leaflet induces a local increase of kB1 molecules and membrane disturbances, which lead to (3) internalization through (i) endocytosis or (ii) internalization by an energy-independent process, whereby the ability to insert and induce tumbling of phospholipids and vesicle-like formation is important for the peptide to cross the lipid bilayer. The bioactive face and the hydrophobic patch are shown in dark and light gray, respectively. Chemistry & Biology 2015 22, 1087-1097DOI: (10.1016/j.chembiol.2015.07.012) Copyright © 2015 Elsevier Ltd Terms and Conditions